Does science need a bitcoin?

What if there was an electronic coin in which it's network were to actually solve real world computational problems relating to science.

Let's say a cure for a disease can only be found using a computer model, they could put a bounty on that cure and when that cure is discovered all the clients running the algorithm would be rewarded with a share of that solution.

Say it takes 1000 blocks to solve a problem worth $1m dollars you get a 1000th of the $1m bounty.

Those sciencecoins would have no value until the solution is found though.